TWINNING IN OTHER SPECIES OF ARMADILLO 8i 



(Fig. 33) in connection with Figs. 12 to 14 for Dasypus. 

 In both, through the process of so-called germ-layer 

 inversion, the ectodermic vesicle is at the distal i)ole of 

 the fixed vesicle and the endodermic vesicle covers the 

 external part of the vesicle from the Trager downwarrl. 

 Evidently then this process of germ-layer inversion is 

 not, as was at first supposed, the key to polyembryonv. 

 It merely furnishes the conditions under which polv- 

 embryonic development may easily occur. Note that 

 in Euphr actus the Trager is in the form of a vesicle with 

 a completely inclosed cavity called the Trager cavity. 

 In D. novemcinctiis there appears to be no true Trager 

 cavity, for the proximal wall, corresponding to that 

 at the top of the Fernandez figures, does not develop. 

 The trophoderm plate is the same in both species, and 

 the thickened ring of trophoderm (the Trager) that in 

 Dasypus invades the maternal mucosa in stages rei^ re- 

 sented in Figs. 12 to 16 corresponds to the thickened 

 side walls of the trophodermic vesicle in Euphractus. 

 The difference in the two species is associated with the 

 fact that in Euphractus the trophodermic vesicle does 

 not sink deeply into the maternal mucosa but lies more 

 on the surface, while in Dasypus the penetration of 

 the Trager into the maternal mucosa is much deeper, 

 involving a practically complete dropping of the proxi- 

 mal wall of the Trager cavity. The sides of this vesicle 

 persist as an ingrowing ring or collar of glandular cells 

 that penetrate rather deeply into the mucosa epithelium ; 

 but the trophodermic plate inclosed by this ring remains 

 free from the mucosa and may sometimes have between 

 it and the underlying mucosa a considerable space which 

 may be homologized with the Trager cavity of Fernandez. 



